Beginning Anomaly Detection Using

training set and testing set, you can use a built-in  

scikit- learn function called train_test_split, as detailed below:

x_train, x_test, y_train, y_test = train_test_split(df2, labels,  

test_size = 0.2, random_state = 42)

The parameters are as follows: x, y, test_size, and random_state. Note that x and 

y are supposed to be the training data and training labels, respectively, with test_size 

indicating the percentage of the data set to be used as test data. random_state is a 

IRU

I

LQ

UDQJH  

GI GILORF>QSUDQGRPSHUPXWDWLRQ OHQ GI @

GI GI>@

ODEHOV GI>ODEHO@

GIBYDOLGDWH GI>@

[BWUDLQ[BWHVW\BWUDLQ\BWHVW WUDLQBWHVWBVSOLW GIODEHOV

WHVWBVL]H UDQGRPBVWDWH 

[BYDO\BYDO  GIBYDOLGDWHGIBYDOLGDWH>ODEHO@

Figure 2-20.  Shuffling the values in df and creating your training, testing, and 

validation data sets

Chapter 2   traditional Methods of anoMaly deteCtion

---

46

number used to initialize the random number generator that determines what data 

entries are chosen for the training data set and for the test data set.

Finally, you delegate the rest of the data to the